Body fatness, in particular, in the abdominal, intra-abdominal and hepatic regions is known to carry a higher metabolic risk, compared to overall, peripheral or subcutaneous fat. Yet, the effect of fat mass in high-risk compartments, which is found in greater proportions in ethnic minorities, is likely to have been underestimated by cancer epidemiology research that relied on proxy measures of body mass and waist size. We hypothesize that body fat topography can be predicted by a combination of demographic, simple anthropometry, and lifestyle exposome variables (exposure-related blood biochemistry and behavior profiles). Indeed, we found in a feasibility study using the Multiethnic Cohort (MEC) study that the prediction of the abdominal fat compartments measured by abdominal MRI can be improved when considering exposome variables in addition to BMI. In the proposed study, we will first assess energetics-related behaviors of dietary intake, physical activity and cognitive eating and will analyze fasting blood levels of adipocytokines, insulin/IGF, sex steroid hormones, and lipids and lipid-soluble micronutrients among 2,000 MEC subjects (200 in each of 10 sex-ethnic groups) who will be recontacted for a cross-sectional study of detailed adiposity phenotypes measured by DXA and abdominal MRI. Using Random Forest and other similar regression models, we will identify the best nutritional, biochemical and behavioral markers that improve the prediction from BMI of the following adiposity phenotypes: total fat mass and trunk-to-periphery fat ratio measured by DXA and deep and superficial abdominal subcutaneous fat, visceral fat and liver fat measured by MRI. Finally, we will test the associations of the predicted adiposity phenotypes, as well as top predictors, for association with the risk of breast (n=1,217 cases) and colorectal cancers (n=1,379 cases) in nested case-control studies using the prospectively collected questionnaire data and biospecimens in the MEC. These data will be essential to this POTs integrated approach to study the relationships of the exposome, genome, metabolome and gut microbiome with body fat distribution and cancer risk since they may help to identify new avenues for behavioral interventions.